INDEX
    Explanations
    New Auto-Interp
    Negative Logits
     invariant
    -0.10
    sharp
    -0.09
     enamel
    -0.08
    ecture
    -0.08
     вд
    -0.08
    gp
    -0.08
    _raise
    -0.08
    -0.08
    -0.08
    -0.07
    POSITIVE LOGITS
     utilized
    0.07
    м
    0.07
    (not
    0.07
    orr
    0.07
     Aeros
    0.07
    ları
    0.07
    0.07
     Matrix
    0.07
     cozinha
    0.07
     Küche
    0.07
    Act Density 0.000%

    No Known Activations